Is There a Role for Machine Learning in Liquid Biopsy for Brain Tumors? A Systematic Review.

The paucity of studies available in the literature on brain tumors demonstrates that liquid biopsy (LB) is not currently applied for central nervous system (CNS) cancers. The purpose of this systematic review focused on the application of machine learning (ML) to LB for brain tumors to provide practical guidance for neurosurgeons to understand the state-of-the-art practices and open challenges. The herein presented study was conducted in accordance with the PRISMA-P (preferred reporting items for systematic review and meta-analysis protocols) guidelines. An online literature search was launched on PubMed/Medline, Scopus, and Web of Science databases using the following query: "((Liquid biopsy) AND (Glioblastoma OR Brain tumor) AND (Machine learning OR Artificial Intelligence))". The last database search was conducted in April 2023. Upon the full-text review, 14 articles were included in the study. These were then divided into two subgroups: those dealing with applications of machine learning to liquid biopsy in the field of brain tumors, which is the main aim of this review (n = 8); and those dealing with applications of machine learning to liquid biopsy in the diagnosis of other tumors (n = 6). Although studies on the application of ML to LB in the field of brain tumors are still in their infancy, the rapid development of new techniques, as evidenced by the increase in publications on the subject in the past two years, may in the future allow for rapid, accurate, and noninvasive analysis of tumor data. Thus making it possible to identify key features in the LB samples that are associated with the presence of a brain tumor. These features could then be used by doctors for disease monitoring and treatment planning.

Paths of Evolution of Progressive Anaplastic Meningiomas: A Clinical and Molecular Pathology Study.

Grade 3 meningiomas are rare malignant tumors that can originate de novo or from the progression of lower grade meningiomas. The molecular bases of anaplasia and progression are poorly known. We aimed to report an institutional series of grade 3 anaplastic meningiomas and to investigate the evolution of molecular profile in progressive cases. Clinical data and pathologic samples were retrospectively collected. VEGF, EGFR, EGFRvIII, PD-L1; and Sox2 expression; MGMT methylation status; and TERT promoter mutation were assessed in paired meningioma samples collected from the same patient before and after progression using immunohistochemistry and PCR. Young age, de novo cases, origin from grade 2 in progressive cases, good clinical status, and unilateral side, were associated with more favorable outcomes. In ten progressive meningiomas, by comparing molecular profile before and after progression, we identified two subgroups of patients, one defined by Sox2 increase, suggesting a stem-like, mesenchymal phenotype, and another defined by EGFRvIII gain, suggesting a committed progenitor, epithelial phenotype. Interestingly, cases with Sox2 increase had a significantly shortened survival compared to those with EGFRvIII gain. PD-L1 increase at progression was also associated with worse prognosis, portending immune escape. We thus identified the key drivers of meningioma progression, which can be exploited for personalized treatments.